Publikationen - Netze für Erneuerbare Energien

Patrick Lehnen, Patrick Mack, Markus de Koster, Ingo Stadler, Eberhard Waffenschmidt,
"Spectral Analysis for Tracking Transient Effects in Distribution Grids",
CIRED-Workshop, Chicaco, 7.-8.11.2024,
IET Conference ProceedingsVol. 2024, Iss. 27, pp 310–314
https://doi.org/10.1049/icp.2024.2622

In the face of the challenges posed by climate change and the rapid integration of renewable energy sources, the resilience of power distribution grids has never been more crucial. Our research introduces a novel methodology aimed at increasing grid resilience by leveraging advanced analytics enabled by simulations. We propose an algorithm to locate failures and power quality issues in the presence of modern high-power, non-linear inverter-based assets like photovoltaics (PV), battery electric vehicle (BEV) charging stations, and heat pumps.
This paper explores the use of neural networks for identifying the injection point of transients and comprises two distinct experiments. Initially, we demonstrate a proof of concept using the amplitude of the third harmonic across all buses in the CIGRE-LV 44-bus distribution grid, achieving a validation accuracy of 100%. Recognizing the impracticality of full observability in real-world scenarios, we strategically reduce measurements to three buses, achieving 84.7% accuracy with a neural network. This method results in top-3 and top-5 categorical accuracies that improve to 97.3% and 98.8%, respectively, giving valuable intuition where the sources might be located.

Downloads

> Preprint: PDF-Dokument (0.5 MB)

> Final Paper, published at IET (The Institution of Engineering and Technology)

Project

> Quirinus-Control